Power Hangs by a Literal Thread

Ian Kumekawa dives into Samanth Subramanian’s “The Web Beneath the Waves: The Fragile Cables That Connect Our World.”

By Ian KumekawaOctober 23, 2025

• Science & Technology

• Politics

NEARLY 20 YEARS AGO, during a congressional debate over net neutrality, Senator Ted Stevens of Alaska described the internet as a “series of tubes.” The remark became an instant meme, a rhetorical relic that suggested an antediluvian age when telecommunications depended on switchboard operators, wires, and cables.

In retrospect, Stevens’s statement was far from absurd. For, as Samanth Subramanian’s excellent new book The Web Beneath the Waves: The Fragile Cables That Connect Our World makes clear, the internet does indeed consist (at least in part) of a vast network of glass tubes—fiber-optic cables. We think of the internet as an abstraction, a view reflected in the lexicon of the data economy—the cloud, AirDrop, even the Ethernet cable. But, though cyberspace may be virtual, it relies on earthly infrastructure. The apparent weightlessness of the internet depends on very physical cables, the most important of which run deep under the world’s oceans. Ninety-nine percent of the world’s data zips through these filaments, which are only three inches wide. These threads on the seafloor are the world’s information superhighways. They are also tremendously fragile, exposed to natural disaster, marine accident, and sabotage. Indeed, the most vulnerable part of the global data infrastructure may well be the part that has been submerged.

Controlling transmission lines of information has long been a source of power. Today’s undersea cables follow the grooves laid down by the British Empire. Submarine telegraph cables were integral to British dominance in the 19th century. By 1907, Britain controlled 75 percent of the world’s underwater telegraphy network. As Subramanian notes, “if you overlaid a modern map of submarine fiber optic networks upon an old map of [British telegraph company] Cable & Wireless cable routes, you’d see astonishingly few differences.” Then as now, cables connected centers of wealth and geopolitical power. The cartography of global power and wealth has changed over the last century, but not unrecognizably.

There are powerful homologies between the first era of undersea cables and the era in which we currently live. Empires understood well the importance of cables. The day after Britain entered World War I, it severed German undersea telegraph lines. After the war, Germany poured huge sums into a network of wireless radio transmission stations as a way to circumvent the British control of undersea telegraphy. Telecommunication was intimately bound up in the geopolitics of the great powers—and still is. Powerful states assert sovereign control over how data is used and where it is stored, but at the same time, as Subramanian points out, “poorer nations find that if they want better internet at all, they must relinquish some of their sovereign control to Western tech giants.”

Those giants are not states but companies: Google, Amazon, Meta, Apple, and others. It is well known that tech giants have controlled a large majority of web traffic for some years. What is less well known is that undersea fiber-optic cables themselves are also increasingly becoming controlled by the world’s most powerful tech companies. Such projects can only be taken on by players with the deepest pockets. As Subramanian shows, 2Africa, the largest and perhaps most important cable system yet built, has been spearheaded and principally funded by Meta at the estimated cost of one billion dollars. A cable envisioned to nearly circumnavigate the globe from the US East Coast to California via South Africa and Australia will run considerably more, and will also be funded by Meta, perhaps without foreign partners. Subramanian predicts that “Google will doubtless consider a rejoinder.”

While today’s tech conglomerates are not official state actors, such companies are just as enmeshed in geopolitical concerns. As Subramanian points out,

neither [Meta nor Google] is likely to sell capacity to Chinese or Russian companies, requiring those countries to lay bigger cables of their own. Not very long ago, it made little sense to speak of “American cables” or “Chinese cables.” But as Meta and Google engirdle the earth in cables, and as governments panic about security, the corporate and nationalist constrictions of the internet will align and overlap.

This will have major implications. In the future, Subramanian claims, “it will become impossible to think of the internet as a single resource.” Already accused of creating a multitude of virtual echo chambers, the internet will also become physically divided along fissures that already split up the spheres of influence of the new great powers: the United States, Russia, and China. Such a future represents a dystopian perversion of the original ambitions of connectivity. It smacks of the tripartite world order outlined in George Orwell’s Nineteen Eighty-Four (1949), a neo-imperial order that is becoming less far-fetched with every passing month.

Powerful states have long relied upon private, profit-seeking actors to carry out their policies, even to carry the flag of empire itself. British colonialism, as historian Philip J. Stern has argued, depended largely on the ministrations and self-interested schemes of corporate actors. Many of these companies—the East India Company, the Hudson’s Bay Company, Cecil Rhodes’s British South Africa Company—were the global economic powerhouses of their day. Corporations like Google and Meta are, in many ways, not too different from their predecessors: they act on their own private motives but simultaneously contribute to US power. Historically, private companies were instrumental in building vital communications infrastructure around the world: telegraph lines, railroads, telephone networks. Often, they worked hand in hand with states (imperial or otherwise) to do so. It was not the British state that built telegraph cables, nor (for most of the 19th century) was it the British state that owned the cables. It was, rather, private companies, most of which merged into the conglomerate Cable & Wireless in the 1920s. But in historical retrospect, it is not unreasonable to compress state and corporate actors, as in the statement “Britain controlled 75 percent of the world’s underwater telegraphy network.”

For Subramanian, there is a difference between the corporate players of today and those of the age of high imperialism. “[N]ever,” he writes, “have these companies been as transnational, wealthy, and concentrated as the tech conglomerates are today.” This is a debatable point—the East India Company controlled a larger standing army in the early 19th century than the British state itself did—but not an unreasonable one. The real question is not size but rather the particular relationship between corporate and state power. The big imperial chartered companies of the 19th century operated with huge leeway but ultimately were both deployed and brought to heel by authorities in London.

To what extent are big tech companies headquartered in Silicon Valley instruments of American power or even American empire? It’s a hard question. As Patrick McGee has recently demonstrated, Apple has become integrally dependent on China. McGee writes that China has “captured” the company. Are big global tech companies really beholden to Western governments the way, for instance, Chinese companies such as Huawei are beholden to Beijing? For decades, there have been fears that technology firms, whether Google, Apple, Meta, or the various concerns of Elon Musk, had become so large and powerful as to be functionally outside the sphere of regulation. There was Apple and Google’s strategic use of the “double Irish” tax-avoidance maneuver. There were claims of anticompetitive and monopolistic behavior, from both right and left.

These fears have not abated, but they have evolved. Big Tech’s moves to please (or set the agenda in) Washington—whether by scrapping DEI initiatives, paying huge sums to do business in China, or presenting ceremonial plaques—signal a willingness to meet political imperatives. One way or another, the US state still has tremendous leverage when it comes to the tech titans. Insofar as Big Tech remains an instrument of American power, cables are too. In fact, Web Beneath the Waves suggests that privately owned physical cables are an integrally important source of geostrategic power today, just as they were for the British Empire.

That power is immense, but it is as fragile as the cables themselves. This is perhaps the key lesson to be learned from focusing on physical infrastructure. Writing in the final years of the Iraq War, political theorist Timothy Mitchell argued that oil production lent itself both to concentration and to physical disruption. Producing oil—compared, for instance, to coal—is tremendously capital-intensive but not terribly labor-intensive. This feature makes it easy for big businesses to dominate the oil trade. But because of the physical geography of oil production—a valuable substance transported vast distances in thin conduits—oil infrastructure is highly subject to physical attack. Oil concentrates power, both in the hands of capitalists (or rulers) and in those of insurgent disrupters. What Mitchell claimed about oil is even truer of fiber-optic data. “For Africa,” Subramanian writes, “oil is indeed the old data.” And not just for Africa—if the world of the 2000s ran on oil, the world of today increasingly runs on data.

Undersea data cables are arguably even more at risk of disruption than oil pipelines are. Subramanian recounts how, in December 2021, an undersea volcano erupted 40 miles off the coast of Tonga’s largest island, severing the single data cable connecting the Pacific island nation to the wider world. In the blink of an eye, the internet went dead, and Tonga was cast into the distant past, back “to a time before the telegraph and scheduled flights reached these parts of the Pacific.” Since Tonga’s telecoms authority had stopped renewing the nation’s subscription to satellite internet and had dismantled its equipment, the aftermath of the cable’s break was like a scene out of a postapocalyptic movie: contact with the outside world was lost; phone connections were broken; medical records were rendered unavailable, bank accounts inaccessible, and packages undeliverable.

While controlling cables is a tremendous source of power, Tonga’s experience is a reminder that such power hangs by a literal thread. Cable networks are huge geopolitical liabilities, far more vulnerable than they were when Britain cut German telegraph lines over a century ago. Subramanian points out that Taiwan is connected to the outside world by 15 cables (as opposed to Tonga’s one), but it suffers from the same structural weakness, one with much larger strategic stakes. Taiwan is not just a key US ally but also a financial hub and the center of the world’s semiconductor industry. Over a three-year span, Taiwan’s cables have experienced more than 50 cuts. Because of seismic activity off the island’s ocean-facing eastern coast, cables make landfall on Taiwan’s western coast, which faces mainland China. A recent incident resulted in two domestic Taiwanese cables breaking. Purportedly, it was an accident, caused by a civilian Chinese fishing boat. The truth of the matter remains murky.

It’s now commonplace to think of cyberspace as a realm of geopolitical contestation. The news is full of organized cyberattacks, state-backed hacking centers, disinformation campaigns, new forms of digital surveillance. But, as Subramanian insists, virtual spaces are always physically tethered. And in an age of power politics, that physical infrastructure frequently serves as a venue of contestation and a source of risk. For when it comes to undersea cables, there is relatively little that even wealthy states can do to prevent disruption. Cables lie on seabeds deep beneath international waters, outside the jurisdiction of individual nations. The international law that governs the high seas—the UN Convention on the Law of the Sea (UNCLOS)—does not include provisions for boarding or inspecting ships suspected of damaging cables, nor does it outline civil liability for cable damage. On or under the open seas, might often makes right. Control of cables is a wellspring of money, power, and influence. Right up until they break.

The Web Beneath the Waves is an elegant study of a hidden world. It also offers a bleak reminder of how far the world has slipped from the heady days of Geocities, MySpace, and AOL. For early digital boosters, the internet was a liberatory, joyful space. It would, so they claimed, bring us closer together and democratize information through a vibrant new public sphere, resistant to domination by the companies that now largely control online life. The optimism of the 1990s and 2000s also figured the World Wide Web as connective tissue—a structure that could transcend national boundaries. Reflecting from the less idealistic vantage point of 2025, Subramanian highlights instead the internet’s potential for literal disconnection, foregrounding how states and companies are leveraging what was once understood as a utopian technology as a tool of realpolitik, reminiscent of the empires of old.

Nothing is more immediate than the internet. But few things feel so physically distant. Attending to the physical dimension of cyberspace is vital, whether to shore up its weaknesses or to exploit them; doing so means thinking about geopolitical influence and control. The stakes are high. After all, as Subramanian writes, “what happened to Tonga could, in theory, happen to anyone—even to the world’s biggest, wealthiest nations.”